[Security Tokens Refinancing] MIP6 Application for OFH Tokens

I’m not a french lawyer either, but in TradFi secured financings a lender usually advances a loan to a borrower who grants a security in favour of the lender over specific assets owned by the borrower itself. The borrower maintains ownership of the asset until an event of default occurs and the security is enforced, resulting in the transfer of the ownership of the asset to the lender.

The overcollateralization bit is not clear to me. The MIP6 Applications indicates that "If the redemption amount paid by SG SFH to the Security Agent exceeds the amount due under the DAI Loan plus the unwinding costs, the Security Agent will transfer the exceeding amount to SG". Thus, MakerDAO would not be entitled to get 100% of the redemption price at maturity, but only an amount equal to the DAI Loan plus unwinding costs.


I could not see how collateral agent is defined like with other participants, given that the collateral agent performs the important role of the collateral test to ensure that the collateral is valued correctly.

Is compensation by SG via SG-Forge to MakerDao community for providing funds to fund their bond position, explicitly stated somewhere?

Hello everyone,

Here is GoldenNaim from sushi, and I want to share my personal opinion.
Firstly, as a person living in France, i’m very proud and happy to see this MIP6 Application from one of the largest bank in France ( top 2 ) , in UE ( top 5 ) and in the world ( top 20 ) .

It shows us how TradFI and DeFI protocol can work together.
More important : I believe that an important part of institutional funds will probably come from the activity of refinancing & loan.

It’s a win-win deal for TradFI & DeFI

A small tech briefing :

The transparency is important for the community. Let’s take a look of the ERC20 OFH Token contract 0x9915CeD2820082a25A8992752AA43bCB0E1d1097 :

  • Even if the source code isn’t open, we can retrieve these data :
    owner : 0x7b66e68c64e051db9876f07c24d093a38389e825
    initialSupply : 400 ( :white_check_mark: )
    currentSupply : 400
    name : VULC01 ( :white_check_mark: )
    symbol : VULC01 ( :white_check_mark: )
    ISIN : FR0013510518 ( :white_check_mark: )
    undefined : 100000 ( NB : 400*100000 = 40,000,000€ - same amount)
    divisor : 100
    startDate : 1588896000 — Friday 8 May 2020 00:00:00
    undefined : 1747180800 — Wednesday 14 May 2025 00:00:00 — Maturity Date ( :white_check_mark: )
    undefined : 1747180800
    undefined : 1620950400 — Friday 14 May 2021 00:00:00 —
    undefined : 12
    undefined : 0
    undefined : 257
    These data correspond with the bond here : Covered Bond Label

  • Even if it’s an ERC20 Token, by reading the potential failed log events, a lot of improvement have been done with, it seems, the goal to match traditional standard for this type of bonds ( whitelisted investors, possibility to lock/unlock the token, impossibility to transfer the token after a date ). List of errors which can be emitted by the contract :

    • ERROR : Only issuer can perform this action ( :white_check_mark: )
      ERROR : Subscription ticket not locked
      ERROR : Transfer balances failed
      ERROR : Issuance has matured. Tokens cannot be transferred anymore ( :white_check_mark: )
      ERROR : Only operator with registrar role can unlock token ( :white_check_mark: )
  • The contract has been created by another contract ( CAST FRAMEWORK ? ) : 0x62eacbd2801ee8ae866ef57909eebbf407dd1dd7
    We can still use the function : getTokenByName(string name) , getAllTokens().

    • getTokenByName(“VULC01”) returns the correct address of the OFH Tokens held by Societe generale : 0x9915CeD2820082a25A8992752AA43bCB0E1d1097 . ( :white_check_mark: )
    • Only a whitelisted address can use the contract, otherwise you the contract will return the error : Calling address should match registrar agent

Short resume of the OFH Tokens / covered bonds :

If I read correctly, from multiple sources, they are 100% backed by home loans.
The cover pool is composed of multiple mortgage, which represent a large over-collateralization , up to 140% ( :white_check_mark: )
Every trimester, a global report can be downloaded here for people who want to know more about what is behind :

These bonds had ( and still have ) a strong AAA from Fitch & Moody, with a stable outlook, official source : ( :white_check_mark: )

OTC Trading // exchange for security tokens

Following the launch of Trident, Sushi will begin formalizing franchised pools for institutional and other permissioned use cases. Franchised pools are a way to allow users to provide liquidity on decentralized exchanges while meeting their compliance needs. As such, these pools will be differentiated from the main Trident AMM system and will allow whitelisting and similar features for liquidity providers and swappers.

Let’s discuss another day about Franchised pools available soon on Sushi.

Risks & Benefits

For this specific DAI Loan, the risk seems extremely low, especially because of the over-collateralized percentage ( the french Law want a minimum of 125% , it’s around 140% ). Otherwise, the liquidation process can be done quickly.

Well, all DeFI’s actor must be happy of this news. The bond refinancing is a big market, with serious institutional investors, who are able to inject hundred of billions inside our eco-system, which is exactly what we want and I’m personally very happy for this potential collaboration between Maker & SG .


This is an exciting transaction - with a top tier counterparty and excellent collateral! And rated AAA.

Here are some thoughts on the structure and risk of the proposed transaction.

I think its important to remember this proposed transaction is a covered bond structure as opposed to a securitization structure. Covered bonds are senior secured debt instruments typically issued by a bank. In addition to the recourse to the issuer a covered bond investor also has a preferential claim to a separate “cover pool” of mortgage loans or other high quality assets meant to be isolated from the issuer in an insolvency. Thus the risk analysis of a covered bond covers the (1) corporate issuer, the (2) “cover pool” of assets and (3) the cover structure that gives investors a preferential claim to the “cover pool” if needed.

This covered bond deal is rated AAA by Fitch (aaa by Moodys).
SocGen is currently rated A- by Fitch
The credit of SocGen and the covered pool together are AAA
Implicitly the covered pool must be rated AAA

When Fitch rates this covered bond AAA that means that it can withstand a AA+ rating stress and still pay out timely principal and interest. Given that a AA+ rating stress would cause Soc Gen (rated A-) to default, this implies that the covered bond must look mainly to the covered pool of assets to pay its timely principal and interest and that the covered bond structure functions as designed. Note that the ratings stress is not just credit focused but also interest rate focused so if the covered pool are interest rate sensitive (such as fixed rate mortgages), they must be able to withstand both a AA+ credit stress and a AA+ interest rate stress.

The covered bonds and the covered pool of French mortgages are denominated in Euros and hence there is not currency risk to the covered bond. However, the MarkerDao loan will be in DAI so the currency risk needs to be addressed.

I have provided links below for
A) the underlying covered pool below (by Fitch)
B) Fitch’s Covered Bond rating criteria
C) A thoughtful discussion on covered bond risks by PIMCO

A) Link to Fitch data on covered asset pool:

B) Link to Fitch Covered Bond Rating Criteria:

C) Here is a good discussion from PIMCO on covered bonds


Some details on the covered asset pool (Fitch data on covered asset pool) as of June 30 2021:
(currency in euros)

42.83 billion in mortgages in pool
60 covered bonds in pool
37.04 billion covered bonds
1.6 billion average covered bond
5.6 year WA maturity in covered bonds
3.15 billion covered bonds with maturity in less than 1 year

350,976 mortgages in pool
122,749 average mortgage (Euros)
99% fixed rate mortgages
46.6 months WA seasoning of mortgages
14.8 year WA maturity of mortgages
68.2% WA CLTV of mortgages
100% guaranteed mortgages
78.3% owned occupied mortgages
0.0% mortgages in arrears
100% French mortgages
36.9% Ile-de-France (incl. Paris)
9.3% Nord-Pas-de-Calais

15.6% overcollateralization (OC)
5.0% minimum regulatory OC%
5.0% minimum contractual OC%
7.0% break even OC%
2.9% break even credit loss OC%
4.2% break even ALM loss OC%


While this is a very positive first step, let’s not forget that we should remain focused on a future where we tokenize all 350k mortgages in the pool for full on-chain transparency. We should see all tranches, not just the AAA tranche on-chain.

A protocol is a system of collaboration and this is a perfect example where multiple parties can collaborate on-chain. The issuer, the credit ratings agencies, the valuation firms, the auditors, the security agents, the exchange agents, the Maker representative, and any investors in the junior tranches should work in unison to establish the risk, get it priced, and then get it funded. Such a structure will help solve some of the audit deficiencies that surfaced in 2008 and it will lead towards removing the friction in the TradFi system that will result in greater access to capital and a lower cost of capital delivered by DeFi. This is the future of finance and the future of DeFi.


This is a great development and congrats to the CU teams.

We are meeting with a U.K. ‘high street’ bank later this week and this would be a great story to share, especially as the initiative is coming a French bank :slight_smile:

Whilst this is inevitable - the DeFi and CeFi interoperability - their choice of partner is a proof point of how competitive Maker can be in terms of funding costs. As MKR holders, we probably won’t get much in terms of direct benefits either as stability fee or at liquidation.

Going with Maker or not, SC has plenty of other funding options eg retail deposits and bond markets. Other borrowers do not have the same luxury, especially those from high growth emerging markets. So hope this sort of deals don’t crowd out other deal where Maker can make a huge difference in terms of financial inclusion and revenue for the protocol too.

Cover bonds do have cover pools which may or may not sit in a separate SPV but all the while should be trackable on chain, ultimately. The memory of 2008 is still fresh and our bank partner was bailed due to the loss of nexus between underlying mortgages and the myriad of instruments that were layered on top of them.


Agreed - all the underlying assets should be on-chain BUT they contain personal info on the mortgage holders that cannot be broadly shared. Is it easy enough to sufficiently anonymize the mortgage personal info to protect the mortgage holders but also include enough info to help the investors?

I don’t think that securing personal info is a real concern for the current tech, but bringing natively on-chain the whole chain will definitely be a true challenge. Especially considering the fact that some authorities and counterparts will have to get comfortable with the lack of central enforcement of certain actions - which is the reason to exist of authorities and the reason not to of DeFi.

Still, incredible step forward. I am working on a piece for Dirt Roads, the newsletter I host, that goes into the details of the deal - my very personal opinions over it. I will share this here in the thread for anybody interested.

1 Like

Data Privacy and Security Law in Europe GDPR requires very strict handling of personal data on Europeans. But agreed with @luca_pro it should not be an issue for current tech. Depending on granularity of the underlying assets it is not even necessary to look at every individual exposures (apart from verifying existence and claims) but to use statistical analyses to assess portfolio risk. This technique is used by rating agencies and is accepted by institutional investors.

1 Like

Dear MakerDAO community,

First, thank you for all your messages. On reading your messages, Iet me clarify/precise the following points.

Purpose of the experimentation:

This first use case is designed to shape our future refinancing offering on Security Tokens. This first step focuses on legal, operational, and technical integration feasibility. To ease the realization of this experimentation:

  • We will use a series of OFH Tokens issued by SG SFH and currently held by SG and for which SG-Forge is the Registrar; and
  • The OFH Tokens are fixed rate covered bonds at 0% in order to avoid asset servicing complexity.

If this first experimentation is conclusive, we would propose this refinancing setup to our clients, i.e. the investors in Security Tokens issuances. At this stage, SG-Forge is structuring debt instruments in the form of Security Tokens covered bonds, Notes (secured and unsecured) and structured products (Notes linked to underlying assets).

Experimentation concerns and scale-up option:

We are aware of the weakness of the liquidation process and collateral valorization in this experimentation framework. We think that below elements mitigate these drawbacks:

  • The covered bond tokens are highly rated.
  • In the context of this experimentation, the DAI Loan will be overcollateralized enough with such a high quality assets. For information purposes, retained covered bonds have a haircut between 8% to 10%. We invite the community to request the ECB collateral eligible assets (Query eligible assets) with comparable OFH bonds issued by SG SFH (e.g. FR0013358496, FR0013232071, FR0013259413). At inception, the DAI Loan will be collateralized at least at 130%. We propose a liquidation ratio at or around [115%]. We will also ensure a minimum operational ratio that will be higher than the liquidation ratio [at or around 120%] (either to add collateral, or to refund DAI or to unwind the structure).
  • It will be a short-term refinancing experiment. We expect to terminate the refinancing structure after a period of 6 to 9 months maximum to be efficient enough.
  • In case of the enforcement of the pledge, the covered bonds will have a short remaining maturity (less than 4Y).
  • Because the OFH Tokens are not liquid, we assume that the price will be close to par. However, the French Law covered bonds are a highly regulated and transparent. I complete the Naïm post (thank you by the way for the clarity of your post) with the following link that provides all the information regarding the SFH covered bonds issuances and the cover pool: Investisseurs dette - Société Générale under “Société de financement de l’habitat” and “HTT Covered Bonds Label Reporting / Asset Reports”.

In a second time, to offer this refinancing solution to security token holders and to comfort MakerDAO, we will work on an efficient liquidation process with appropriate market data oracles for this kind of assets. Those features should be assessed post experimentation. For example, several initiatives could be looked at like a network of keeper for Security Token or a SushiSwap facility for Security Token (as suggested by Naïm). Also, the European regulation is under progress to enable Security Token exchanges (i.e. Pilot Regime Regulation).

Parameters of the experimentation:

  • Maturity of the refinancing: from 6 to 9 months.
  • Stability fees: stability fees should be close to the refinancing rate (repo rate) of a covered bond asset plus a liquidity premium.
  • Liquidation ratio: we propose at or around [115%]. An internal operational threshold at or around [120%] will be monitored to avoid liquidation.

OFH Token Smart Contract:

The smart contract relies on CAST that complies with legal and regulatory frameworks. For that purpose, only whitelisted operators can interact with the smart contract.

We can of course make a workshop with MakerDAO core units to explain the CAST framework and to discuss the technical integration. But I think it will be the case if the greenlight is granted.

Thank you,



An truly exciting proposal from SG; this is just another example to show that yields generated by off-chain real world asset (RWA) is the future destinatino for DeFi liquidity. Although DeFi TVL has exceeded $100bn over the past 18month, such number is still dwarfed by the size of real world finance, which is measured in $trillions.

Me and my team have also launched a project called EntroFi to connect DeFi capitals directly with loan borrowers who use real world assets (account receivables, real estate properties, pledges of securities, etc.) or existing NFTs as collaterals. We envision a world where liquidity and assets alike flow freely between on chain and off chain, and therefore would love to work with anyone team or platform who share such vision with us.

This could be something truly extraordinary.

1 Like

That’s Awesome, looking forward to witnessing this EntroFi project! :money_mouth_face: :money_mouth_face: :money_mouth_face:

I believe that this project is setting the bar for the quality of the interactions that intermediaries should and will have with the Maker ecosystem, independently from the outcome of the collateral onboarding process.

I am looking forward to read more from the teams which are DD’ing the deal on what is the community’s reaction to this proposal.


i particularly like the diagram that includes monies flows back and forth; in fact; I have gone ahead and implemented a similar one to our application which I will be submitting promptly.
A question however remains open whether MakerDAO is open to use a Security Agent and if so, to onboard more than 1 proposed in the application.

Dear SG-Forge,

The CurioDAO, a collaborator with MakerDAO, proposes a unique solution that could perhaps help alleviate some of the weaknesses of the liquidation process.

We are of the opinion, utility tokens and security tokens are converging, and provided our solution supports this collab, I would be more than happy to do an intro with our legal consul respectively share more about this solution.

By bringing additional liquidity to security tokens via a novel Wrapped token module via a token that represents solely the ownership to the underlying asset is not considered as a security.

For more about this topic, see option 2 in CurioDAO’s legal opinion

[Analysis on wrapped token_20210428_final.pdf - Google Drive](https://4.- Wrapping Token Legal Note)

More about CurioDAO here

Hi, I’m new to this, and I haven’t fully understood the MakerDao whitepaper, so these questions may be dumb. As a lawyer think I fairly understand the proposal though. Anyhow:

  1. How is the collateral test made to ensure the loan is sufficiently collateralized ? Are the OFH tokens or their underlying assets publicly traded anywhere ?

  2. Why is the entity making the collateral test (the Security agent) the same as the Maker representative ? Why was it chosen by SG and what legal obligations will it have towards individual members of the community as a representative ?

  3. In the event that the collateral crashes in value, won’t the value of the emitted DAI crash as well ? Would that not give SG 5 business days to buy DAI back at a very low value with its USD and reimburse its loan ?

Love the initiative. Just wondering who’s taking most of the risk in this transaction.

1 Like

DAI generated by OFH is not only backed by OFH. DAI is backed by the total value locked in the protocol plus the market cap of MKR. If OFH becomes worthless then any shortfall will be made up by selling MKR.


Going to vote yes for this assuming DD will be completed.

A lot of questions. One of them is regarding the DC here. I understand this being an experiment sounds like SF is going to be 0 here and it looks like the DC is going to be roughly full value to CR of 130% which seems somewhat odd for an experimental deployment.

Supportive of experiments but my issue is containing risk on these things to something the protocol can handle (i.e. SB on hand which is around 50M).

I also get concerned when I heard the following from the RWF CU lead.

Even so I see no reason to stop this from next stages and expect relevant CUs to step forward with analysis and commentary here vs. just leaving this for governance to go/no go. We really need good color from people who are being paid to do these jobs (risk, RWF, and if PECO has anything).

I agree with others this is exciting if Maker can properly get this under its belts.

With so much going on I have some concerns of over work and some critical issues falling through cracks due to how these are being pressed forward. It was my hope we would have something of a standard model for RWF before bringing on so many new RFW players.